IIT Home Page CNR Home Page

Numerical Methods for Large Problems

This activity is aimed at the analysis and development of algorithms for the solution of very large linear systems (i.e., 10,000 equations and 10,000 variables). These problems are often associated to queue problems or arise in the discretization of infinite domain differential equations. A second topic concerns the study of iterative methods for ill-conditioned problems, like in some problems of image reconstruction. A third reseach activity concerns the analysis of the stability properties of “fast” methods for the solution of structured linear systems.

05-2009

Research theme: Algorithms and Computational Mathematics

Partecipants:

Representative: Paola Favati

Foto di Paola Favati
Publication
Dario Bini, Paola Favati, Beatrice Meini
A compressed cyclic reduction for QBDs with low rank upper and lower transitions
2011, The Seventh International Conference on Matrix-Analytic Methods in Stochastic Models (MAM7)
Franco Brezzi, Dario Bini, Sebastiano Seatzu, Gianfranco Capriz, Bruno Codenotti, Paola Favati, editori
Calcolo. a Quarterly on Numerical Analysis and Theory of Computation, Vol. 48 n.1,n.2, n.3, n.4
2011
Paola Favati, Grazia Lotti, Ornella Menchi, Francesco Romani
Stopping rules for iterative methods in nonnegatively constrained deconvolution
2011
Paola Favati, Grazia Lotti, Ornella Menchi
Superfast solution of Toeplitz-like systems
2011
Favati P., Lotti G., Menchi O., Romani F.
Performance analysis of Maximum Likelihood methods for nonnegative image deconvolution
2010, Inverse Problems
Paola Favati, Grazia Lotti, Ornella Menchi
Recursive Algorithms for Unbalanced Banded Toeplitz Systems
2009, Numer. Linear Algebra Appl.
Paola Favati , Grazia Lotti, Ornella Menchi
Stability of Levinson algorithm for Toeplitz-like Systems
2009